Stem Cells from the Mammalian Blastocyst


  • Janet Rossant Ph.D.

    Corresponding author
    1. Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Ontario, Canada and Department of Molecular and Medical Genetics, University of Toronto, Ontario, Canada
    • Samuel Lunenfeld Research Institute, Mount Sinai Hospital, 600 University Avenue, Toronto, Ontario, Canada M5G 1×5. Telephone: 416-586-8267; Fax: 416-586-8588
    Search for more papers by this author


Early differentiation of the mammalian embryo leads to the development of two distinct lineages—the inner cell mass (ICM) and the trophectoderm. Cells of the ICM are pluripotent and give rise to all tissues of the fetus, while trophectoderm cells are restricted in their potential to the trophoblast cell layers of the placenta. In the mouse, apparently immortal stem cell lines can be obtained from both cell types. These cell lines, embryonic stem (ES) cells and trophoblast stem (TS) cells, are morphologically and molecularly distinct and depend on different signaling pathways for their maintenance. They also show different cell fates when introduced into early embryos to generate chimeras. However, a change in the levels of expression of a key regulator of pluripotency, Oct4, can push ES cells towards the TS phenotype, when grown in TS cell conditions. Stem cell potential in the early embryo thus appears to depend on a combination of the levels of expression of key intrinsic regulators and the appropriate extrinsic environmental factors. Manipulation of both intrinsic and extrinsic regulators may be needed to reveal the full potential of stem cells from other stages of development and the adult.


There is currently considerable excitement and controversy about the potential use of pluripotent stem cells as a source of differentiated cells for repair of degenerating or damaged tissue in humans. It has been well established in the mouse that stem cells with the capacity to differentiate into all tissue types of the adult can be derived from the early embryo. Harnessing the amazing potential of these embryonic stem (ES) cells and channeling their differentiation down specialized pathways needed for tissue-based repair are active areas of research. Recent studies have reported successful derivation of multiple cell types, including neurons, glia, neuronal stem cells, islet cells, hepatocytes, osteoblasts, and adipocytes from mouse ES cells in vitro [1]. This research has gained momentum from the reports that cells with similar properties to mouse ES cells can be derived from human blastocysts [2] or fetal gonads [3], thus bringing the goal of ES cell-based cell therapy closer to fruition.

Human ES cell research is controversial, however, mostly because of the source of material used to generate the lines. Because of this and the intrinsic advantages of accessing adult tissues, there is also intensive research on the potential ability of tissue-restricted adult stem cells to be redirected into other cell lineages. The classic view of development is a progressive restriction of cell fate from the totipotent zygote to the highly specialized differentiated tissues of the adult. Stem cell development has been thought of in the same manner. Self-renewing stem cells exist and can be isolated at different stages of development. Earlier stem cells, such as ES cells, are likely to be pluripotent, while stem cells derived from specific tissues later in development or the adult, such as hematopoietic or neural stem cells, are expected to be restricted in their potential to the tissue from which they arise. This paradigm is currently under challenge, with increasing evidence that some adult tissue-restricted stem cells may contribute to other cell types when exposed to the appropriate environmental influences [4–, 8]. Perhaps most striking have been reports that neural stem cells derived from the adult brain have the capacity to contribute extensively to multiple cell types in vivo when reintroduced into the early embryo [9]. The definition of a pluripotent versus a restricted stem cell is thus becoming somewhat blurred, and the possibility that stem cell restriction versus plasticity is determined solely by extrinsic environmental influences becomes a plausible hypothesis.

Here I describe the properties of two different kinds of stem cells that derive from the mouse blastocyst—the ES cell and the trophoblast stem (TS) cell. These two different kinds of stem cells derived from the very early embryo use very different signaling pathways to maintain stem cell proliferation, require different transcription factor expression for specifying the stem cell state, and contribute to entirely different lineages when injected into the supposedly permissive environment of the mouse blastocyst. Thus, they behave as restricted stem cells in terms of in vivo cell fate. However, ES cells can be switched towards the trophoblast lineage by changing culture conditions and altering expression levels of Oct4, a key determinant of pluripotency. I suggest that this model may have relevance to adult stem cell plasticity. A combination of environmental influences plus alteration in expression levels of a few key regulatory genes may impart restricted stem cells with a much broader range of developmental potential than is initially apparent.

Lineage Specification in the Blastocyst

At embryonic day 3.5 in the mouse, the embryo undergoes its first overt differentiation into the inner cell mass (ICM) and the trophectoderm. The trophectoderm is an epithelial monolayer of cells, enclosing a fluid-filled cavity, the blastocoel, at one end of which is a compact group of cells, the ICM. Many experimental lineage studies have shown that the ICM gives rise to the embryo itself plus extraembryonic membranes like the allantois and the amnion, while the trophectoderm contributes solely to the trophoblast layers of the placenta [10]. Prior to implantation, a new cell layer differentiates on the blastocoelic surface of the ICM, the primitive endoderm, which is also an entirely extraembryonic cell type, giving rise to the endoderm layers of the yolk sacs [11]. The trophectoderm differentiates further before implantation: the trophectoderm overlying the ICM continues to proliferate and will go on to form the diploid trophoblast of the placenta, while the cells away from the ICM cease division but continue to endoreduplicate their DNA to form the trophoblast giant cells. Morphological and molecular analysis has shown that all three early lineages are distinct by the late blastocyst stage.

The molecular events underlying the separation of these distinct cell lineages are still not well understood. However, it is known that the POU-domain transcription factor Oct4 (Pou5f1), is essential for development of the ICM. Oct4 is expressed throughout oogenesis and preimplantation development, becomes restricted to the ICM at the blastocyst stage and later is expressed throughout the early epiblast, before being confined to the developing germ cells [12, 13]. It is thus a key marker of cells with pluripotentiality, and understanding its regulation and its targets is critical for understanding the pluripotent state. Oct4 mutant embryos initially resemble blastocysts but express no ICM markers and can only generate trophoblast cells in vitro [14]. Thus, Oct4 is required for ICM development. Identification of key positive regulators for specification of the trophectoderm lineage has been less successful. There are several transcription factors, such as the bHLH gene Mash2 [15], the caudal-related gene Cdx2 [16], and the T-box gene Eomes [17], whose expression pattern in early development is the converse of Oct4, i.e., they are expressed throughout the early embryo but become restricted to the trophectoderm at the blastocyst stage. However, mutation of these genes individually does not result in a failure to specify trophectoderm. Mash2 is involved in later placental development [18], but Cdx2 and Eomes are required for early trophoblast development. Both mutants block development at the blastocyst/peri-implantation stage and show failure of trophoblast proliferation [19,, 20]. Trophectoderm formation is initiated, however, suggesting that neither alone is required for initial trophectoderm specification. The phenotype of double mutants will be of interest in this regard.

Although information is limited, it is clear that the transcription factor hierarchy governing cell fate, as well as the downstream differentiation targets, are already very different in the ICM and the trophectoderm by the late blastocyst stage.

Stem Cells from the Blastocyst

ES Cells

During development in vivo, the ICM forms the early postimplantation epiblast, which is a transient pluripotent pool of cells that rapidly differentiates at gastrulation into the primary germ layers of the developing fetus. Although the presence of self-renewing, pluripotent stem cells is transient in vivo, apparently immortal cell lines with these properties can be obtained in vitro by culture of ICM cells of the blastocyst in the presence of the cytokine, leukemia inhibitory factor (LIF) [21,, 22]. These ES cells can be maintained indefinitely in the presence of LIF, and express markers of the undifferentiated, pluripotent state, including Oct4. Upon removal of LIF, they downregulate markers like Oct4, rapidly lose self-renewal capacity and differentiate into a variety of cell types. Generation of chimeras by aggregation of ES cells with eight-cell embryos or injection of ES cells into blastocysts revealed that ES cells are pluripotent but not totipotent. Although they contribute to all tissues of the fetus, they fail to contribute to the trophectoderm or primitive endoderm lineages [23]. This failure is the basis of the tetraploid complementation assay, in which totally ES-derived fetuses can be obtained by aggregation of ES cells with tetraploid embryos [24, 25]. Tetraploid embryos do not generate embryonic lineages but can provide trophectoderm and primitive endoderm to support embryonic development from the ES cells.

The pathway by which LIF signaling acts to promote ES cell self-renewal has been well studied, largely by Austin Smith and colleagues. LIF signals via heterodimerization of the two class I cytokine receptors, the low-affinity LIF receptor (LIF-R), and the common subunit, gp130 [26]. Homodimerization of gp130 alone is sufficient to drive ES cell proliferation without involvement of the LIF-R subunit [27]. Fusion constructs of the extracellular domain of the G-CSF receptor (G-CSF-R) with the transmembrane and cytoplasmic domain of gp130 were transfected into LIF-R-deficient ES cells, generating lines in which ES cell proliferation was promoted in the presence of G-CSF rather than LIF [28]. This approach allowed the assessment of the relative roles of different signaling pathways downstream of gp130 in ES cell self-renewal. The cytoplasmic domain of gp130 contains several tyrosine residues that are phosphorylated by associated JAK kinases after ligand-stimulated dimerization. Four of these phosphorylated tyrosines have been identified as putative interaction sites with the SH2 domains of the transcription factor, STAT3 [29]. Mutation of the four STAT binding sites in the G-CSF-R-gp130 fusion construct was shown to abrogate STAT3 activation and block ES cell self-renewal. A dominant negative STAT3 construct showed the same effect [28].

Stimulation of gp130 signaling in ES cells also phosphorylates SHP-2 and leads to activation of the mitogen-activated protein (MAP) kinases, ERK1 and ERK2 [30]. However, mutation of Y118, the SHP-2 binding site in gp130 [31], did not block ES cell proliferation, nor did incubation of ES cells in a specific inhibitor of MAP kinase activation [32]. In fact, inhibition of the SHP-2-RAS-ERK pathway actually seemed to promote self-renewal over differentiation. Although gp130 signaling can activate MAP kinase, such activation is more commonly associated with signaling downstream of receptor tyrosine kinases (RTKs). In embryogenesis RTK signaling is critical for the differentiation and morphogenesis events of gastrulation. Thus, blocking MAP kinase activation in ES cells may act to block the differentiation pathways promoted by MAP kinase signaling and indirectly enhance stem cell maintenance.

These data clearly support a necessary role for the gp130-JAK-STAT3 pathway in ES self-renewal. STAT3 activation is also sufficient for stem cell maintenance [33]. The in vivo relevance of the LIF pathway for embryo development is not entirely clear. LIF is expressed in the trophectoderm of the blastocyst and the LIF receptor in the ICM, as would be expected if this pathway is required for pluripotent cell survival in vivo [34]. However, neither LIF mutants [35] nor mutants of the receptors, LIF-R [36, 37], and gp130 [38], show any defects in the development of the ICM or early epiblast. Recent evidence suggests that the LIF pathway is necessary for survival of the ICM during implantational delay [39], suggesting that evolution has co-opted this pathway to support a specialized aspect of rodent reproduction. This function was then fortuitously diverted to allow the derivation of ES cells.

TS Cells

The critical dependence of ES cells on a balance between activating STAT signaling while repressing MAP kinase signaling is born out by the fact that a relatively small change in growth factor conditions promotes the isolation of an entirely different stem cell from blastocyst outgrowths—the TS cell. TS cells can be derived from blastocysts or early postimplantation trophoblast by culture in the presence of fibroblast growth factor (FGF)4 plus heparin and primary embryo fibroblast-conditioned medium [40]. In these conditions, ES cells do not develop but indefinitely self-renewing diploid epitheloid cell lines can be isolated. These cell lines express markers typical of early diploid trophoblast and are fully dependent on FGF signaling for their proliferation. Removal of FGF promotes differentiation of the cells towards the giant cell fate. Addition of LIF has no effect on either proliferation or differentiation. Like ES cells, TS cells also retain the capacity to contribute to normal tissues when reintroduced into the early embryo [40]. However, TS cells only contribute to the trophoblast lineages of the placenta and not to the fetus itself. They behave as restricted stem cells, capable of recapitulating the entire differentiation pathway of the trophoblast lineage, but not able to make any other cell types under the conditions tested.

TS cells require MAP kinase activation for their proliferation. The signal transduction events downstream of ligand binding and FGF receptor activation are complex and variable in different spatial and temporal milieux. However, receptor binding and phosphorylation of the docking protein, FRS-2, has been shown to be a key event, since FRS-2 couples FGF signaling to the MAP kinase pathway [41,, 42]. Examination of the biochemical changes in TS cells following FGF4 activation has revealed phosphorylation of key components of the FRS2-RAS-MAP kinase pathway. Such activation is necessary for stem cell maintenance, since treatment with a MAP kinase inhibitor blocks self-renewal and promotes differentiation to giant cells (L. Corson, unpublished observations).

FGF signaling appears to play a critical role in trophoblast development in vivo, as well as in vitro. FGF4 is expressed in the ICM and early epiblast [43,, 44], while the receptor, FGFR2, is strongly expressed in the overlying trophoblast [45]. Mutants of both genes have been reported to arrest in development around implantation, consistent with a failure of trophoblast proliferation [46,, 47]. MAP kinase activation, as revealed by immunostaining with antibody to phosphorylated ERK1 and 2, occurs in trophoblast cells adjacent to the early epiblast (L. Corson, unpublished observations), as predicted if FGF signaling is critical for stem cell proliferation. The direct targets of FGF signaling in the trophoblast in vivo and TS cells in vitro are not yet clear. However, the two transcription factor genes, Cdx2 and Eomes, that were mentioned as being required for early trophoblast lineage development are potential FGF targets. Their expression pattern coincides with the zone of MAP kinase activation in the early postimplantation embryo, and they are expressed in TS cells and rapidly downregulated upon FGF removal.

TS cells provide a novel system for examining the pathways controlling proliferation and differentiation in the trophoblast lineage with some reasonable expectation that findings in TS cells will be relevant to events in vivo. TS cells may also retain some of the specialized properties of the mammalian extraembryonic lineages, such as preferential paternal X-inactivation, and DNA methylation status, providing new tools for examining the molecular underpinnings of these events.

ES Cells to TS Cells—A Simple Switch?

Culture conditions that activate different intracellular signaling pathways thus lead to the isolation and maintenance of two very different stem cell populations from the mouse blastocyst. Activation of the JAK-STAT pathway and inhibition of MAP kinase signaling promotes self-renewal of a pluripotent stem cell with properties similar to the early epiblast and expressing a key regulator, Oct4. Activation of MAP kinase signaling through FGF receptor activation promotes proliferation of a restricted stem cell from the trophoblast lineage, which expresses key regulators, such as Cdx2 and Eomes. These two cell lines are morphologically and molecularly distinct, their potential in vivo is very different, and they use different pathways to promote self-renewal. Interestingly, despite being derived from the early embryo, where cells might be expected to be redirected fairly easily into other pathways, ES cells never make trophoblast and TS cells never make embryonic tissues under any conditions tested so far in vivo or in vitro.

Although environmental conditions promoting a switch between the two cell types have not been found, a genetic switch that allows ES cells to reveal TS potential in TS culture conditions has been determined. Niwa and colleagues [48] developed ES cell lines in which Oct4 could be conditionally repressed and showed that such lines failed to remain as ES cells, even in the presence of LIF, but differentiated into cells resembling trophoblast giant cells. If culture conditions were switched from ES to TS conditions [40], downregulation of Oct4 occurred, resulting in derivation of proliferating cell lines showing all the properties of TS cells. Thus, downregulation of Oct4 is necessary for trophoblast development in vitro and probably in vivo in the blastocyst as well. Clearly, ES cells have all the machinery in place to undergo trophoblast differentiation and only require the flip of a genetic switch—Oct4 OFF—and the appropriate growth factor stimulation, to undertake an entirely different pathway of development (Fig. 1).

Figure Figure 1..

Interconvertibility of stem cells from the mouse blastocyst.

The formula of Oct4 ON = ES, Oct4 OFF = TS cells may be too simplistic. It is currently unknown whether TS cells can be switched to the ES fate by turning Oct4 on constitutively. However, the general concept that there may be key genetic switches that can open up new developmental pathways for stem cells in culture is a useful one. Altering environmental conditions in which stem cells are grown can have considerable effects on their developmental potential. However, if tissue-restricted stem cells express key transcriptional regulators that inactivate entire genetic pathways, extracellular signaling events alone may not be sufficient to reactivate the pathway. If the key intrinsic regulators can be identified, their expression can be manipulated genetically and new stem cell potentials revealed. This concept of combining genetic regulation and environmental manipulation to reveal full stem cell potential will be important for stem cell therapeutics in the future.

Stem Cells from the Human Blastocyst—Do the Same Rules Apply?

The signaling pathways and downstream transcription factors involved in the derivation and maintenance of mouse blastocyst-derived stem cells are often assumed to be conserved across mammalian species. However, it has proved extremely difficult to derive bona fide ES cells from many mammalian species, using related approaches to those used for deriving mouse ES cells. There is less experience currently in the derivation of TS cells across species, but we have not yet been able to readily derive human TS cells under similar conditions to those used for deriving mouse TS cells. The derivation of ES-like cells from primate [49] and then human blastocysts [2] was, therefore, an important breakthrough. However, the signaling pathways required for self-renewal of these cells are currently unclear. It has been reported that human ES cells are LIF-independent and that FGF is used in the culture medium in which they are grown [50]. If the cells are truly FGF rather than LIF-dependent, this would be more similar to the conditions used to derive TS cells in mice. This implies that human ES cells are not identical to mouse ES cells and may have some features in common with TS cells. Human ES cells can differentiate into many different cell types in vitro, including perhaps trophoblast, since they express the placental marker, hCG [51]. Further characterization of the signaling pathways required for human ES cell derivation and maintenance will help define the relationships between the different stem cells derived from human and mouse blastocysts.


I thank Tilo Kunath and Laura Corson for useful comments on the manuscript. The author is a Distinguished Scientist of the Canadian Institutes for Health Research and an International Scholar of the Howard Hughes Medical Institute.